What it Looks Like: Cardiac Arrest and CPR

Update: Our friends at EMS 12 Lead have put together a “sister post” to this one, with further discussion and some additional clips. Check it out!


Although we’ve talked about the fundamentals of good CPR before (and then again), the fact remains that the first step of any resuscitation is recognizing the presence of cardiac arrest. In fact, failure to do this in a timely fashion is a common problem at all levels of healthcare: because these situations don’t happen often, we are reluctant to accept when they’re happening now. (Real emergencies don’t come heralded by a change in soundtrack.) The result is delays, often for many minutes, before anybody initiates CPR and attempts defibrillation. We can’t just point fingers at the bystanders and lay providers — it’s also the EMTs, the nurses, even the doctors doing this. “Is that a pulse?” we muse. “I think there’s a pulse. Here, come feel.”

It’s true that cardiac arrest, at least in the early stages, is often not easily distinguished from other maladies (such as unconsciousness due to seizure or drugs). A few clues may be immediately obvious, such as pallor of the skin if some time has passed, or if a bystander actually witnesses the patient suddenly collapse. However, in the end, the way to make this call quickly and reliably is to simply follow the algorithm. You’re not the first person to deal with this, and the American Heart Association has spent years simplifying the decision process — because the goal isn’t to eventually “figure it out,” the idea is to immediately recognize it and start lifesaving measures within seconds.

Is the patient responsive? (No; they appear unconscious, and make no response whatsoever to painful stimuli.) Are they breathing normally? (No; they’re not breathing, or merely performing agonal, “gasping” breaths.) Is there a carotid pulse? (No, no pulse is palpable within a few seconds.) That’s good enough for us. Start pushing on their chest and don’t stop unless it’s absolutely essential — and the only things that are absolutely essential are checking their cardiac rhythm (just a few seconds) and delivering a shock (less than a second).

We’re going to look at a number of examples of real-life cardiac arrest (or “codes” in the usual lingo). As a rule, the actual CPR that you’ll see here is of relatively poor quality. This is due to a number of factors, but primarily it’s because 1) Many of these clips are five, ten, or fifteen years old, from a time when CPR was taught and practiced differently; and 2) Even today, many people do not perform good CPR.

So: focus on the patients. Watch how they present, their breathing, their skin, their responses to the interventions. Watch the challenges that the providers face as far as managing the patient and the environment. Watch how their approaches differ by region, circumstance, or personal preference. But for the most part, do not do what they are doing. We’ll watch a couple examples of really good CPR at the end so you know what to strive for.


We’ve linked this before, and for good reason; it’s one of the best videos I know of a real code. This is older CPR, with less emphasis on compressions and more on ventilation, but otherwise fairly true to the textbook. Notice the early “activation” of EMS, and the brief pulse check. Notice how rather than trying to “one-man” the BVM, they take advantage of the many available hands, allowing one person to hold the mask and one to squeeze the bag. Notice how they quickly dry the chest for the AED without being obsessive about it. As for the compressions, nowadays we would like to see them faster and deeper, with fewer and briefer pauses.

In the patient, watch the spastic, gulping movements of the mouth and tongue; this is agonal breathing. Notice also the decorticate posturing of the upper body, suggesting neurological dysfunction. Finally, notice how (after the third round of CPR + defibrillation), he begins to breathe spontaneously, with obvious chest rise, and this is clearly different from the prior agonal respirations.


(watch through 8:45) Despite the numerous pauses for commentary, this is also good. The initial compressions are rapid — a little too rapid, which is okay, but not deep enough, and if they were deeper they would likely be at a more reasonable rate. The second compressor goes deeper, but does not recoil fully at the top. The third (male) rescuer gives perhaps the best compressions, but notice his elbows — although pushing hard and deep, he allows his elbows to bend slightly each time. This is a very common error in otherwise skilled compressors, and is a good way to fatigue yourself quickly. Make a conscious effort to lock the elbows out completely, allowing you throw your full weight behind each compression rather than “pressing” with the arms. Notice also how frequently the rescuers stop compressions for one reason or another. Chest compressions need to build upon each other for several compressions before you’re producing anything like the coronary perfusion pressures you want to see; repeatedly stopping and starting sacrifices all your hard work.

In the patient, notice the pallor (paleness) of his skin, and the total lack of tone (limp flaccidity) of his body. Notice how he convulses with the shock, and how his chest rises and expands with ventilations. Finally, notice how his abdomen recoils outward in a seesaw manner with each downward compression of the chest.


(watch through 7:10) This is a chest pain patient that codes on camera. Despite the low image quality, notice how poorly he immediately presents; he is obviously fatigued, wan, and struggling with some sort of pain or other internal distress. Upon attempting to stand, he loses consciousness and demonstrates agonal respirations (listen to the heavy snoring). They ask if he has a history of seizures; a substantial number of cardiac arrests are initially mistaken for seizures, and may present with seizure-like activity (such as foaming of the mouth). There is obvious difficulty with compressions due to the high position of the stretcher. Bubba was very fortunate to arrest in the immediate presence of paramedics.


(watch through 3:43) Notice again the initial hesitation due to bystanders believing a seizure is occurring. These compressions have the kind of violent depth we want, although at about half the rate. Notice again the slight arm bend.


A chest pain patient who deteriorates into a full arrest while on camera for a UK documentary. Depicts a good portion of the code.


[Added 5/8/13 — ed.]

(watch until the credits)

ED footage of EMS bringing in a code. Shows the practice of “code surfing,” where a rescuer rides the stretcher to provide ongoing compressions during movement — a great idea if you can do it safely and effectively (it helps to use someone small!) Notice how fast some of the compressions are performed, but it’s tough to reach good depth at those rates, particularly when the arms aren’t held straight. Although the captions note that the patient had ROSC, it’s extremely unlikely that he survived to discharge; when patients are transported without achieving ROSC in the field, they almost never walk out of the hospital. Cardiac arrests are worked on scene; transport without a pulse is simply giving up, unless you have good reason to think there’s a reversible etiology of arrest that the hospital can address.


[Added 8/21/12 — ed.]

(watch through 12:05, or stay for some bystander interviews) Another near-drowning. Decent-looking compressions and a reasonable attempt to minimize interruptions, although notice the pauses for intubation and at various other times. Unknown outcome.


(watch through 2:25) This is a volunteer crew from AMR’s disaster response team in Haiti. There seems to be initial confusion about whether the patient is pulseless or merely apneic, hence the initial focus is on the airway; nowadays we would frown upon interrupting compressions for intubation, and the bagging after the tube has been placed is far too fast (every 6-8 seconds only, please). The teamwork is good, and return of spontaneous circulation (ROSC) is achieved after a few minutes. Notice the decision to defer a blood pressure measurement, since the patient has a strong radial pulse — an indicator of a decent pressure, if not an exact number. The patient does have fixed and dilated pupils, indicating a probable poor neurological status.

Keep watching only if desired; the patient is transported to the field hospital, where she rearrests, and the doctor there halts resuscitation efforts.


(watch through 23:50) This is a neonatal resuscitation immediately following a field delivery of twins; one infant is apneic following birth. BVM ventilations and compressions are performed, as well as an aborted attempt at intubation; however, in the end the neonatal fundamentals of warming, suctioning, stimulation, and supplemental oxygen end up effectively reviving the child.



[will not embed; click through to view video then return] Another infant resuscitation, this one in the ED. Excellent footage of compressions, ventilation, and the typical hubbub of a code, as well as an IO (intraosseous) line that infiltrates and the use of ultrasound to assess for cardiac function during PEA.


CPR on a near-drowning. A fine example of the typical poor quality of bystander compressions; notice the negligible depth and general uncertainty about whether to intervene.


A collapse at a sporting event. There is no backstory available on this, so it may not be a true arrest, but if so it would be consistent with commotio cordis, when a blow to the chest (such as a punch) causes an arrhythmia (due to an R-on-T induced by the physical blow; this is the evil brother of a precordial thump, with the opposite effect). This type of arrest has extremely good prognosis for recovery if immediate CPR and defibrillation is performed, since there may be little to no underlying disease; it’s a healthy young patient who simply got whacked wrong.


(watch through :38) Some brief miscellaneous footage of an arrest post-drowning, with a few pretty good compressions.


(watch through :57) Another near-drowning. Nice compressions. Notice the pallor and lack of tone.


[Added 10/11/13 — ed.]

This is clearly an old video, although it’s not clear from what year. Regardless, it’s a great opportunity to list the things you’d do differently today. Since we know that the keys to a successful resuscitation are immediate, deep, fast, uninterrupted compressions, along with rapid defibrillation, do you think this patient had a good outcome? How many of the interventions they performed instead of that stuff are still recommended care? If you were on that scene, would you be an advocate (some might say a CPR Nazi) to ensure that things were done properly?


Finally, let’s look at a couple examples of really spot-on, perfect resuscitation. Since perfection is rare in life, and having a camera in the room is even rarer, these will be simulations.

Click here for a teaching video from the Austin/Travis County medical director’s office. It demonstrates their “pit crew” model, where each member has a designated role, and each action is carefully crafted to match the latest evidence for best practices to promote survival. Notice how compressions begin almost immediately, once the rescuers have noted a lack of responsiveness, breathing, and pulse — and compressions stop for almost nothing, no matter what else is happening. (I would call these compressions very good, but a bit fast and shallow.) Secondary tasks like bagging can happen in the background. This crew does stop compressions while the AED charges, while I personally prefer to compress during this interval (between analysis and shock); the longer you delay between last compression and delivery of the shock, the less chance of getting a pulse back.


(Watch from 2:45 onward) This is the model from Salt Lake City Fire, portraying a highly progressive model. Aside from the general concepts of “compression-centered” resuscitation and the pit crew model, they’re also eliminating pauses for rhythm analysis (using the “see-through” filter on the Zoll monitors, which removes CPR artifact) and even for defibrillation (shocking without taking hands off the chest, which has not been proven safe, but generally seems to be). In other words, there’s essentially no interruption in compressions until there’s evidence of a perfusing rhythm. Notice the compression technique, where knuckles remain against the chest to lock-in the hand position, but the heel of the palm comes off at the top, ensuring full recoil. Beautiful stuff.


There you have it, folks: what dead people look like, and what it looks like when we try to bring them back. Typically the process is chaotic, and we do our best, but often drop the ball on what’s important. Nobody’s perfect, but we can direct our focus toward the pieces that matter the most, and this lets us “streamline” our efforts away from the distractions and toward the critical elements. Recognize the problem early, compress hard, deep, and fast, and don’t stop for anything unless it’s defibrillation. Ain’t so hard, is it?


Sincere thanks to James Oz (Melclin) for assistance with compiling these video clips.


Check out also what Jugular Venous DistentionSeizures, and Agonal Respirations look like


  1. If you look closely in the Seattle CPR example, you’ll see David Baumrind from the EMS 12-Lead Blog video taping their high performance CPR example too.

  2. Great collection of videos! Thanks!

  3. Here’s one that was just posted recently.


  4. Where did the Seattle video go?

  5. Would you continue compressions if a heart beat is picked up by doppler?

    • Interesting question, jasmin. The advent of easily-available ultrasound technology (whether simple dopplers or full duplex-capable portable ultrasound) has made the concept of “PEA” complicated. One can have no palpable pulse but have a dopplerable one (or visible cardiac activity on echo).

      I think most people have decided that if the large pulses are dopplerable only (femoral, carotid), but not palpable, the patient is indeed in PEA (or “pseudo PEA”) for practical purposes. But there’s room to disagree.


  1. […] good friend Brandon Oto over at EMS Basics. He’s been gracious enough to allow us to adapt the original post from his What it Looks Like series over here. We highly suggest that you go check out the amazing […]

  2. […] also what Agonal Respirations, Seizures, and Cardiac Arrest and CPR look […]

  3. […] friend Brandon Oto over at EMS Basics. He’s been gracious enough to allow us to adapt the original post from his What it Looks Like series over here. We highly suggest that you go check out the amazing […]

Speak Your Mind